A configurable online reputation aggregation system

Abstract

Online reputation systems are currently receiving increased attention while online interactions are flourishing. However, they lack one important feature: globality. Users are allowed to build a reputation within one online community, and sometimes several reputations within several independent online communities, but each reputation is only valid within the corresponding community. Moreover, such reputation is usually aggregated by the provider of the online reputation system, giving the querying agent no say in the process. This thesis presents a novel solution to this problem. We conduct a literature review on existing trust and reputation models and classify these models using proper criteria. We introduce an online reputation system that collects reputation information about a ratee from several online communities and allows for this information to be aggregated according to the inquiring agent's own requirements. We propose a configurable aggregation method for local and global reputation based on a discrete statistical model, taking into account several factors and parameters that qualify the reputation. We also implement a prototype of the proposed reputation computation model.